Development and validation of a radiomic signature to predict HPV (p16) status from standard CT imaging: a multicenter study

Ralph T. H. Leijenaar, Marta Bogowicz, Arthur Jochems, Frank J. P. Hoebers, Frederik W. R. Wesseling, Sophie H. Huang, Biu Chan, John N. Waldron, Brian O'Sullivan, Derek Rietveld, C. Rene Leemans, Ruud H. Brakenhoff, Oliver Riesterer, Stephanie Tanadini-Lang, Matthias Guckenberger, Kristian Ikenberg, Philippe Lambin

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Objectives: Human papillomavirus (HPV) positive oropharyngeal cancer (oropharyngeal squamous cell carcinoma, OPSCC) is biologically and clinically different from HPV negative OPSCC. Here, we evaluate the use of a radiomic approach to identify the HPV status of OPSCC. Methods: Four independent cohorts, totaling 778 OPSCC patients with HPV determined by p16 were collected. We randomly assigned 80% of all data for model training (N = 628) and 20% for validation (N = 150). On the pre-treatment CT images, 902 radiomic features were calculated from the gross tumor volume. Multivariable modeling was performed using least absolute shrinkage and selection operator. To assess the impact of CT artifacts in predicting HPV (p16), a model was developed on all training data (Mall) and on the artifact-free subset of training data (Mno art). Models were validated on all validation data (Vall), and the subgroups with (Vart) and without (Vno art) artifacts. Kaplan-Meier survival analysis was performed to compare HPV status based on p16 and radiomic model predictions. Results: The area under the receiver operator curve for Mall and Mno art ranged between 0.70 and 0.80 and was not significantly different for all validation data sets. There was a consistent and significant split between survival curves with HPV status determined by p16 [p = 0.007; hazard ratio (HR): 0.46], Mall (p = 0.036; HR: 0.55) and Mno art (p = 0.027; HR: 0.49). Conclusion: This study provides proof of concept that molecular information can be derived from standard medical images and shows potential for radiomics as imaging biomarker of HPV status. Advances in knowledge: Radiomics has the potential to identify clinically relevant molecular phenotypes.
Original languageEnglish
JournalBritish Journal of Radiology
Volume91
Issue number1086
DOIs
Publication statusPublished - 2018

Cite this

Leijenaar, R. T. H., Bogowicz, M., Jochems, A., Hoebers, F. J. P., Wesseling, F. W. R., Huang, S. H., ... Lambin, P. (2018). Development and validation of a radiomic signature to predict HPV (p16) status from standard CT imaging: a multicenter study. British Journal of Radiology, 91(1086). https://doi.org/10.1259/bjr.20170498
Leijenaar, Ralph T. H. ; Bogowicz, Marta ; Jochems, Arthur ; Hoebers, Frank J. P. ; Wesseling, Frederik W. R. ; Huang, Sophie H. ; Chan, Biu ; Waldron, John N. ; O'Sullivan, Brian ; Rietveld, Derek ; Leemans, C. Rene ; Brakenhoff, Ruud H. ; Riesterer, Oliver ; Tanadini-Lang, Stephanie ; Guckenberger, Matthias ; Ikenberg, Kristian ; Lambin, Philippe. / Development and validation of a radiomic signature to predict HPV (p16) status from standard CT imaging: a multicenter study. In: British Journal of Radiology. 2018 ; Vol. 91, No. 1086.
@article{448d28bfff5145a0bdd0201cb1d78a44,
title = "Development and validation of a radiomic signature to predict HPV (p16) status from standard CT imaging: a multicenter study",
abstract = "Objectives: Human papillomavirus (HPV) positive oropharyngeal cancer (oropharyngeal squamous cell carcinoma, OPSCC) is biologically and clinically different from HPV negative OPSCC. Here, we evaluate the use of a radiomic approach to identify the HPV status of OPSCC. Methods: Four independent cohorts, totaling 778 OPSCC patients with HPV determined by p16 were collected. We randomly assigned 80{\%} of all data for model training (N = 628) and 20{\%} for validation (N = 150). On the pre-treatment CT images, 902 radiomic features were calculated from the gross tumor volume. Multivariable modeling was performed using least absolute shrinkage and selection operator. To assess the impact of CT artifacts in predicting HPV (p16), a model was developed on all training data (Mall) and on the artifact-free subset of training data (Mno art). Models were validated on all validation data (Vall), and the subgroups with (Vart) and without (Vno art) artifacts. Kaplan-Meier survival analysis was performed to compare HPV status based on p16 and radiomic model predictions. Results: The area under the receiver operator curve for Mall and Mno art ranged between 0.70 and 0.80 and was not significantly different for all validation data sets. There was a consistent and significant split between survival curves with HPV status determined by p16 [p = 0.007; hazard ratio (HR): 0.46], Mall (p = 0.036; HR: 0.55) and Mno art (p = 0.027; HR: 0.49). Conclusion: This study provides proof of concept that molecular information can be derived from standard medical images and shows potential for radiomics as imaging biomarker of HPV status. Advances in knowledge: Radiomics has the potential to identify clinically relevant molecular phenotypes.",
author = "Leijenaar, {Ralph T. H.} and Marta Bogowicz and Arthur Jochems and Hoebers, {Frank J. P.} and Wesseling, {Frederik W. R.} and Huang, {Sophie H.} and Biu Chan and Waldron, {John N.} and Brian O'Sullivan and Derek Rietveld and Leemans, {C. Rene} and Brakenhoff, {Ruud H.} and Oliver Riesterer and Stephanie Tanadini-Lang and Matthias Guckenberger and Kristian Ikenberg and Philippe Lambin",
year = "2018",
doi = "10.1259/bjr.20170498",
language = "English",
volume = "91",
journal = "British Journal of Radiology",
issn = "0007-1285",
publisher = "British Institute of Radiology",
number = "1086",

}

Leijenaar, RTH, Bogowicz, M, Jochems, A, Hoebers, FJP, Wesseling, FWR, Huang, SH, Chan, B, Waldron, JN, O'Sullivan, B, Rietveld, D, Leemans, CR, Brakenhoff, RH, Riesterer, O, Tanadini-Lang, S, Guckenberger, M, Ikenberg, K & Lambin, P 2018, 'Development and validation of a radiomic signature to predict HPV (p16) status from standard CT imaging: a multicenter study' British Journal of Radiology, vol. 91, no. 1086. https://doi.org/10.1259/bjr.20170498

Development and validation of a radiomic signature to predict HPV (p16) status from standard CT imaging: a multicenter study. / Leijenaar, Ralph T. H.; Bogowicz, Marta; Jochems, Arthur; Hoebers, Frank J. P.; Wesseling, Frederik W. R.; Huang, Sophie H.; Chan, Biu; Waldron, John N.; O'Sullivan, Brian; Rietveld, Derek; Leemans, C. Rene; Brakenhoff, Ruud H.; Riesterer, Oliver; Tanadini-Lang, Stephanie; Guckenberger, Matthias; Ikenberg, Kristian; Lambin, Philippe.

In: British Journal of Radiology, Vol. 91, No. 1086, 2018.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Development and validation of a radiomic signature to predict HPV (p16) status from standard CT imaging: a multicenter study

AU - Leijenaar, Ralph T. H.

AU - Bogowicz, Marta

AU - Jochems, Arthur

AU - Hoebers, Frank J. P.

AU - Wesseling, Frederik W. R.

AU - Huang, Sophie H.

AU - Chan, Biu

AU - Waldron, John N.

AU - O'Sullivan, Brian

AU - Rietveld, Derek

AU - Leemans, C. Rene

AU - Brakenhoff, Ruud H.

AU - Riesterer, Oliver

AU - Tanadini-Lang, Stephanie

AU - Guckenberger, Matthias

AU - Ikenberg, Kristian

AU - Lambin, Philippe

PY - 2018

Y1 - 2018

N2 - Objectives: Human papillomavirus (HPV) positive oropharyngeal cancer (oropharyngeal squamous cell carcinoma, OPSCC) is biologically and clinically different from HPV negative OPSCC. Here, we evaluate the use of a radiomic approach to identify the HPV status of OPSCC. Methods: Four independent cohorts, totaling 778 OPSCC patients with HPV determined by p16 were collected. We randomly assigned 80% of all data for model training (N = 628) and 20% for validation (N = 150). On the pre-treatment CT images, 902 radiomic features were calculated from the gross tumor volume. Multivariable modeling was performed using least absolute shrinkage and selection operator. To assess the impact of CT artifacts in predicting HPV (p16), a model was developed on all training data (Mall) and on the artifact-free subset of training data (Mno art). Models were validated on all validation data (Vall), and the subgroups with (Vart) and without (Vno art) artifacts. Kaplan-Meier survival analysis was performed to compare HPV status based on p16 and radiomic model predictions. Results: The area under the receiver operator curve for Mall and Mno art ranged between 0.70 and 0.80 and was not significantly different for all validation data sets. There was a consistent and significant split between survival curves with HPV status determined by p16 [p = 0.007; hazard ratio (HR): 0.46], Mall (p = 0.036; HR: 0.55) and Mno art (p = 0.027; HR: 0.49). Conclusion: This study provides proof of concept that molecular information can be derived from standard medical images and shows potential for radiomics as imaging biomarker of HPV status. Advances in knowledge: Radiomics has the potential to identify clinically relevant molecular phenotypes.

AB - Objectives: Human papillomavirus (HPV) positive oropharyngeal cancer (oropharyngeal squamous cell carcinoma, OPSCC) is biologically and clinically different from HPV negative OPSCC. Here, we evaluate the use of a radiomic approach to identify the HPV status of OPSCC. Methods: Four independent cohorts, totaling 778 OPSCC patients with HPV determined by p16 were collected. We randomly assigned 80% of all data for model training (N = 628) and 20% for validation (N = 150). On the pre-treatment CT images, 902 radiomic features were calculated from the gross tumor volume. Multivariable modeling was performed using least absolute shrinkage and selection operator. To assess the impact of CT artifacts in predicting HPV (p16), a model was developed on all training data (Mall) and on the artifact-free subset of training data (Mno art). Models were validated on all validation data (Vall), and the subgroups with (Vart) and without (Vno art) artifacts. Kaplan-Meier survival analysis was performed to compare HPV status based on p16 and radiomic model predictions. Results: The area under the receiver operator curve for Mall and Mno art ranged between 0.70 and 0.80 and was not significantly different for all validation data sets. There was a consistent and significant split between survival curves with HPV status determined by p16 [p = 0.007; hazard ratio (HR): 0.46], Mall (p = 0.036; HR: 0.55) and Mno art (p = 0.027; HR: 0.49). Conclusion: This study provides proof of concept that molecular information can be derived from standard medical images and shows potential for radiomics as imaging biomarker of HPV status. Advances in knowledge: Radiomics has the potential to identify clinically relevant molecular phenotypes.

UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85047962422&origin=inward

UR - https://www.ncbi.nlm.nih.gov/pubmed/29451412

U2 - 10.1259/bjr.20170498

DO - 10.1259/bjr.20170498

M3 - Article

VL - 91

JO - British Journal of Radiology

JF - British Journal of Radiology

SN - 0007-1285

IS - 1086

ER -